Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Algorithmic Principles

Author: David Dory

Date: 03:19:53 03/13/03

Go up one level in this thread


On March 13, 2003 at 03:59:13, Steven Chu wrote:

>I am doing some research on different search algorithms, in particular:
>alpha beta pruning
>minimax
>minimax with alpha beta cutoff
>negamax
>negamax with alpha beta cutoff
>quiescence search
>
>I just wanted to know if anyone has any ideas on the comparisons and contrasts
>of each of these different algorithmic processes compared with eachother and
>humans.  Also if anyone knows anywhere where i can obtain such information i
>would be most grateful.
>
>Thank you
>
>Steven Chu

Steven,
Just click on Computer Resource Center, and then click on Programming ...

There is a great wealth of data, a Google search will be very profitable, as
well.

Minimax is very useful in two player games of perfect information (ie., all
knowledge is available to both players, like chess), but it involves searching
thru EVERY leaf of the search tree - much too laborious and time consuming even
with fast computers.

Alpha-Beta gives the exact same info, while visiting far fewer leaves of the
search tree, by pruning off those leaves which can not possibly improve the
score for the player. With proper move ordering (best moves first), the savings
is very significant. Without good move ordering, the savings can be zip.

Quiescence as I use the term is not a search, it is a state of the position on
the board, as opposed to dynamic. When quiescence is judged true for a position,
an evaluation is done, score returned, and no more searching is made beyond that
position, on that branch of the tree. That position is now the search horizon or
leaf node, for that branch of the game tree.

Naturally, humans play chess very differently than computers. After years of
trying to "make the computer play chess like a man", it has been found most
effective to let the computer play chess the best way it can - like only a
computer would play. Indeed, despite several studies, psychologists are unable
to quantify how grandmasters skillfully apply their knowledge to the game.

If you have a specific question, please ask away. If you need very broad info on
the algorithmic principles used by computers to play chess - please read some of
the info on the net.

Good luck
Dave






This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.